Permalink
Switch branches/tags
Nothing to show
Find file Copy path
Fetching contributors…
Cannot retrieve contributors at this time
306 lines (234 sloc) 11.8 KB
---
title: "The Causes of Recurrent Conflict"
subtitle: POSC 3610 -- International Conflict
author: Steven V. Miller
institute: Department of Political Science
titlegraphic: /Dropbox/teaching/clemson-academic.png
date:
fontsize: 10pt
output:
beamer_presentation:
template: ~/Dropbox/miscelanea/svm-r-markdown-templates/svm-latex-beamer.tex
latex_engine: xelatex
dev: cairo_pdf
fig_caption: false
slide_level: 3
make149: true
mainfont: "Open Sans"
titlefont: "Titillium Web"
---
```{r setup, include=FALSE}
knitr::opts_chunk$set(cache=FALSE)
library(tidyverse)
library(stevemisc)
library(countrycode)
library(knitr)
library(maddison)
library(stringr)
library(dotwhisker)
NDY <- read_csv("~/Dropbox/projects/mid-project/gml-mid-data/2.03/gml-ndy-disputes-2.03.csv")
MIDA <- read_csv("~/Dropbox/projects/mid-project/gml-mid-data/2.03/gml-mida-2.03.csv")
```
# Introduction
### Puzzle for Today
*What explains the duration of peace following a conflict?*
# Conflict Recurrence
### The Problem
All conflicts must end (Ikle, 1971) but disputes routinely recur.
- France and Prussia/Germany contested 12 MIDs between 1848-1941.
- Four were wars and an additional six were uses of force.
- India and Pakistan have been in 39 MIDs between 1947 and 2010.
- Only four were anything other than uses of force or war.
- Likewise, Israel and Syria have been in 34 MIDs between 1948 and 2010.
###
```{r, eval=F, echo=F, fig.width=14, fig.height=8.5, warning=F, message=F}
tibble(year = seq(1947, 2010),
const = 1) -> cyear
NDY %>% filter(ccode1 == 750 & ccode2 == 770) %>%
select(dispnum, year, fatality) %>%
group_by(year) %>%
summarize(sum = n(),fatality = max(fatality)) %>%
full_join(., cyear) %>%
arrange(year) %>%
mutate(sum = ifelse(is.na(sum), 0, sum)) %>%
ggplot(.,aes(year,sum)) + theme_steve() +
geom_bar(stat = "identity", fill = "black", color="black", alpha=I(0.6)) +
scale_x_continuous(breaks = seq(1945, 2010, by = 5)) +
xlab("Year") + ylab("Number of Ongoing MIDs in the Year") +
annotate("rect", xmin=1946.5, xmax=1951.5, ymin=0, ymax=3, alpha=.3, fill="red") +
annotate("rect", xmin=1954.5, xmax=1956.5, ymin=0, ymax=3, alpha=.3, fill="red") +
annotate("rect", xmin=1957.5, xmax=1959.5, ymin=0, ymax=3, alpha=.3, fill="red") +
annotate("rect", xmin=1961.5, xmax=1962.5, ymin=0, ymax=3, alpha=.3, fill="red") +
annotate("rect", xmin=1963.5, xmax=1967.5, ymin=0, ymax=3, alpha=.3, fill="red") +
annotate("rect", xmin=1970.5, xmax=1972.5, ymin=0, ymax=3, alpha=.3, fill="red") +
annotate("rect", xmin=1980.5, xmax=1981.5, ymin=0, ymax=3, alpha=.3, fill="red") +
annotate("rect", xmin=1982.5, xmax=1986.5, ymin=0, ymax=3, alpha=.3, fill="red") +
annotate("rect", xmin=1989.5, xmax=1991.5, ymin=0, ymax=3, alpha=.3, fill="red") +
annotate("rect", xmin=1992.5, xmax=2003.5, ymin=0, ymax=3, alpha=.3, fill="red") +
annotate("rect", xmin=2007.5, xmax=2010.5, ymin=0, ymax=3, alpha=.3, fill="red") +
labs(title = "India and Pakistan Have Contested 39 MIDs since Independence and Have Been in Almost Constant Conflict",
subtitle = "The red shades capture fatal MIDs, including MIDs that became wars from 1947-1949, 1964-1966, 1971, and 1993-1999.",
caption = "Data: GML MID Data (v. 2.01)")
```
```{r, eval=T, echo=F, fig.width=14, fig.height=8.5, warning=F, message=F}
tibble(year = seq(1947, 2010),
const = 1) -> cyear
NDY %>% filter(ccode1 == 750 & ccode2 == 770) %>%
select(dispnum, year, fatality) %>%
group_by(year) %>%
summarize(sum = n(),fatality = max(fatality)) %>%
mutate(isfatal = ifelse(fatality == 0, 0, 1)) %>%
full_join(., cyear) %>%
arrange(year) %>%
mutate(sum = ifelse(is.na(sum), 0, sum),
isfatal = ifelse(is.na(isfatal), 0, isfatal),
`Fatal MID?` = ifelse(isfatal == 0, "Not Fatal", "Fatal")) %>%
ggplot(.,aes(year,sum, fill = `Fatal MID?`)) + theme_steve_web() +
geom_bar(stat = "identity", color="black", alpha=I(0.8)) +
scale_x_continuous(breaks = seq(1945, 2010, by = 5)) +
xlab("Year") + ylab("Number of Ongoing MIDs in the Year") +
scale_fill_manual(values=c("#bb0000", "gray66")) +
labs(title = "India and Pakistan Have Contested 39 MIDs Since Independence and Have Been in Almost Constant Conflict",
subtitle = "The red shades capture years in which there were fatal MIDs, including MIDs that became wars from 1947-1949, 1964-1966, 1971, and 1993-1999.",
caption = "Data: GML MID Data (v. 2.03)")
```
## Four Primary Approaches
### Four Primary Approaches
1. Bargaining
2. Enduring rivalries
3. Conflict management
4. Deterrence
### Bargaining Approaches to Conflict Recurrence
Bargaining models do well to endogenize (sic) all phases of conflict (see: Filson and Werner). Core findings:
- Power is central to explaining international conflict
- See also: asymmetric information
- Incentives to renegotiate (however arrived) explain conflict recurrence.
- Third-party mediation mollifies commitment problems.
### Enduring Rivalries and Conflict Recurrence
Intuitively, most conflict recurrence is between enduring rivals.
- e.g. India-Pakistan, France-Prussia/Germany, etc.
- Rivalry scholars tend to treat this as a sample selection issue.
There is some focus on how dispute outcomes facilitate new dispute onset.
- Hensel (1994): 93% of disputes followed by a decisive outcome are followed by a new dispute (in Latin America).
### Conflict Management and Conflict Recurrence
Conflict management scholars place high emphasis on how disputes end.
- e.g. third-party intervention, UN peacekeeping missions, even regime type.
- Generally, more multilateralism and more negotiation -> more stability
Some core findings:
- Third-party interventions are good for peace.
- UN PKOs -> more stability after disputes.
Scholars additionally quibble about selection issues.
- e.g. do PKOs and mediators follow the more difficult conflicts or the easier ones?
### Conflict Recurrence as Deterrence Failure?
Scholars in this vein see conflict recurrence as a form of deterrence failure. Findings:
- Imposed settlements are more peaceful than negotiated ones.
- Cease-fires can help maintain peace.
- Victor-imposed regime change are followed by longer peace spells.
## Settlement and Recurrent Conflict
### "Peace for Our Time"
![Neville Chamberlain announces "peace for our time" after arriving from Munich.](MunichAgreement.jpg)
### Announcing Unconditional Surrender Ultimatum in Casablanca
![Allied leaders announce the unconditional surrender policy in Casablanca](casablanca-conference.jpg)
### Types of Settlement
- **Negotiated**: an agreement, formal or informal, struck by both sides of a conflict absent any attempt of external imposition.
- **Imposed**: an agreement forced on another state without invitation.
- **None**: a dispute ends because both sides stop fighting without any agreement about the pre-war status quo or conflict.
###
```{r, eval=T, echo=F, fig.width=14, fig.height=8.5, warning=F, message=F}
MIDA %>%
group_by(settle) %>%
summarize(sum = n()) %>%
filter(settle != -9) %>%
mutate(cat = c("Negotiated", "Imposed", "None", "Unclear"),
perc = round((sum/sum(sum))*100, 2),
perc = paste0(perc,"%")) %>%
ggplot(.,aes(cat, sum)) + theme_steve_web() +
geom_bar(stat="identity", fill = "black", color="black", alpha=I(0.6)) +
geom_text( aes(label=perc), vjust=-.5, colour="black",
position=position_dodge(.9), size=4) +
xlab("Settlement Type") + ylab("Number of MIDs") +
labs(title = "Most MIDs (67%) End in No Settlement Type",
subtitle = "Most MIDs stop because both sides stop fighting and not necessarily that they ''agreed'' to stop fighting.",
caption = "Data: GML MID Data (v. 2.01)")
```
### Why Should Imposed Settlements Be More Peaceful?
Quackenbush argues imposed settlements are a case of unilateral deterrence.
- In mutual deterrence, both sides need to be deterred.
- In unilateral deterrence, only one side (i.e. the loser) needs to be deterred.
Unilateral deterrence is actually more peaceful than mutual deterrence.
###
| **Settlement Type** | **Events** | **E(Events)** | **Chi-sq** | **p-Value** |
|:--------------------|:----------:|:-------------:|:----------:|:-----------:|
| Imposed | 237 | 353.08 | 46.84 | <.001 |
| No Settlement | 1466 | 1341.87 | 31.82 | <.001 |
| Negotiated | 427 | 435 | .19 | .663 |
Table: Log-Rank Tests for Settlement Types and Peace Duration (Quackenbush, 2014)
Notice: we observe fewer peace failures if the settlement in the previous dispute was imposed.
- We observe *more* peace failures if there was no settlement type whatsoever.
###
```{r, eval=T, echo=F, fig.width=14, fig.height=8.5, warning=F, message=F}
tribble(
~term, ~estimate, ~std.error, ~model,
"Negotiated Settlement", .361, .087, "Model 2",
"No Settlement", .513, .087, "Model 2",
"Minimum Democracy", -.020, .006,"Model 2",
"Change in Relative Power", 1.782, .603,"Model 2",
# "Change in Relative Power*ln(Time)", 1.046, .840,
"Contiguity", .775, .079, "Model 2",
"Decisive Outcome", .144, .061, "Model 2",
# "Decisivie Outcome*ln(Time)", -.039, .056,
"War", -.105, .099, "Model 2",
"Negotiated Settlement", .352, .087, "Model 3",
"No Settlement", .507, .087, "Model 3",
"Minimum Democracy", -.020, .006,"Model 3",
"Change in Relative Power", -1.774, 1.63,"Model 3",
"Change in Relative Power*ln(Time)", 1.046, .840, "Model 3",
"Contiguity", .776, .080, "Model 3",
"Decisive Outcome", .144, .061, "Model 3",
"Decisive Outcome*ln(Time)", -.039, .056, "Model 3",
"War", .933, .435, "Model 3",
"War*ln(Time)", -.281, .113, "Model 3"
) %>%
mutate(statistic = estimate/std.error,
p.value = 1.96*pnorm(-abs(statistic))) %>%
dwplot(.,dot_args = list(aes(colour = model, shape = model), size = 1.1)) + theme_steve_web() +
xlab("Coefficient Estimate") +
geom_vline(xintercept = 0, colour = "grey60", linetype = 2) +
theme(legend.position = "bottom") +
labs(title = "Cox Model Results for the Effect of Settlement Type on Peace Duration",
subtitle = "Notice how negotiated/no settlements increase the likelihood of survival ''failure'' relative to the baseline of imposed settlements.",
caption = "Reproduction of Model 2 and Model 3 in Table 13.2 (Quackenbush, 2015)") +
scale_colour_discrete(name = "Model") + scale_shape_discrete(name = "Model")
```
### Empirical Findings
Imposed settlements least likely to be followed by peace failure than other settlement types.
- Follows implications of perfect deterrence theory on unilateral deterrence.
Other findings:
- Jointly democratic disputes have longer peace spells.
- Changes in relative power -> peace failure.
- This importantly goes away the longer the peace spell.
- Contiguous states have shorter peace spells than states further away.
- No differences between war and MIDs.
- Until you model an interaction with time.
###
```{r, eval=T, echo=F, fig.width=14, fig.height=8.5, warning=F, message=F}
tibble(x = seq(0, 350),
`Imposed` = x^-.12,
`Negotiated` = x^-.17,
`None` = x^-.20) %>%
gather(`Settlement Type`, value, -x) %>%
ggplot(.,aes(x,value,linetype=`Settlement Type`, color = `Settlement Type`)) + theme_steve_web() +
xlab("Duration in Months") + ylab("Survival Function") + geom_line(size = 1.5) +
theme(legend.position = "bottom") +
scale_x_continuous(breaks = seq(0, 350, by = 50)) +
ylim(0,1) +
labs(title = "Survival Curves by Settlement Type",
subtitle = "Quackenbush's argument is settlement imposition is a form of unilateral deterrence, which is more peaceful than settlement negotiation.")
```
# Conclusion
### Conclusion
Why do disputes recur?
- Four primary theoretical perspectives
- Bargaining models assume changes in relative power drive conflict recurrence.
- Deterrence models see conflict recurrence as deterrence failure.
- Counterintuitively: settlement imposition works better than settlement negotiation.